Artificial Neural Network Modeling to Predict the Moisture Removal Rate of a Desiccant Liquid Dehumidifier System

Abstract

In this study, an artificial neural network model is developed in order to predict the moister removal rate by a desiccant liquid dehumidifier system. This last can be used for air conditioning; it consists mainly of a dehumidifier, a regenerator, and a liquid desiccant. Using a MATLAB® environment, the developed neural network model is based on the multilayer perceptron that includes an input layer, a hidden layer, and an output layer. The network input parameters are ambient air temperature, liquid desiccant temperature, air flow rate, liquid flow rate, air humidity ratio, and liquid concentration. The network output includes one variable which is the moister removal rate. The values predicted by the model are in a good agreement with the experimental data, with mean square error (MSE) of 0.0849 and correlation coefficient (R) of 0.985for all datasets. The used neural network provides a high accuracy optimization and reliability of the method to predict the performance of a liquid desiccant dehumidifier.